Response Feature Technology for High-Frequency Electronics. Optimization, Modeling, and Design Automation

✍ Scribed by Anna Pietrenko-Dabrowska, Slawomir Koziel

• Publisher: Springer
• Year: 2023
• Tongue: English
• Leaves: 604
• Category: Library

✦ Synopsis

This book discusses response feature technology and its applications to modeling, optimization, and computer-aided design of high-frequency structures including antenna and microwave components. By exploring the specific structure of the system outputs, feature-based approaches facilitate simulation-driven design procedures, both in terms of improving their computational efficiency and reliability. These benefits are associated with the weakly nonlinear relationship between feature point coordinates and design variables, which―in the context of optimization―leads to inherent regularization of the objective functions. The book provides an overview of the subject, a definition and extraction of characteristic points, and feature-based design problem reformulation. It also outlines a number of numerical algorithms developed to handle local, global, and multi-criterial design, surrogate modeling, as well as uncertainty quantification. The discussed frameworks are extensively illustrated using examples of real microwave and antenna structures, along with numerous design cases. Introductory material on simulation-driven design, numerical optimization, as well as behavioral and physics-based surrogate modeling is also included. The book will be useful for readers working in the area of high-frequency electronics, including microwave engineering, antenna design, microwave photonics, magnetism and especially those who utilize electromagnetic (EM) simulation models in their daily routines.

✦ Table of Contents

Preface
Acknowledgments
Contents
Chapter 1: Introduction
References
Chapter 2: Simulation-Driven Design of High-Frequency Structures
2.1 Overview. Notation and Terminology
2.2 Local and Global Parameter Tuning
2.3 Multi-objective Optimization
2.4 Surrogate Modeling
2.5 Uncertainty Quantification
References
Chapter 3: Fundamentals of Numerical Optimization
3.1 Optimization Problem Formulation
3.2 Gradient-Based Optimization
3.2.1 Optimization by Descent Methods
3.2.2 Newton and Quasi-Newton Methods
3.2.3 Qualitative Comparison of Descent Methods
3.2.4 Remarks on Constrained Optimization
3.3 Derivative-Free Optimization
3.3.1 Pattern Search Methods
3.3.2 Hooke-Jeeves Direct Search
3.3.3 Nelder-Mead Algorithm
3.4 Global Optimization: Nature-Inspired Population-Based Algorithms
3.4.1 Fundamentals of Population-Based Metaheuristics
3.4.2 Evolution Strategies
3.4.3 Genetic Algorithms
3.4.3.1 Algorithm Flow and Representation
3.4.3.2 Crossover
3.4.3.3 Mutation
3.4.3.4 Selection
3.4.3.5 Elitism
3.4.3.6 Selected Topics
3.4.4 Evolutionary Algorithms
3.4.5 Particle Swarm Optimization
3.4.6 Differential Evolution
3.4.7 Firefly Algorithm
3.4.8 Other Methods
3.5 Summary
References
Chapter 4: Fundamentals of Surrogate Modeling and Surrogate-Assisted Optimization
4.1 Surrogate-Assisted Optimization
4.2 Surrogate Modeling: Data-Driven Surrogates
4.2.1 Modeling Flow for Data-Driven Surrogates
4.2.2 Design of Experiments
4.2.3 Data-Driven Modeling Techniques
4.2.3.1 Polynomial Regression
4.2.3.2 Radial Basis Functions
4.2.3.3 Kriging
4.2.3.4 Artificial Neural Networks
4.2.3.5 Support Vector Regression
4.2.3.6 Other Approximation Methods
4.2.4 Model Validation
4.3 Surrogate Modeling: Physics-Based Surrogates
4.4 Optimization with Data-Driven Surrogates
4.4.1 Optimization by Means of Response Surfaces
4.4.2 Sequential Approximate Optimization
4.4.3 SBO with Kriging Surrogates: Exploration Versus Exploitation
4.4.4 Summary
4.5 SBO Using Physics-Based Surrogates
4.5.1 Space Mapping
4.5.2 Approximation Model Management Optimization (AMMO)
4.5.3 Manifold Mapping (MM)
4.5.4 Shape-Preserving Response Prediction (SPRP)
4.5.5 Adaptive Response Scaling
4.5.6 Summary
References
Chapter 5: Basics of Response Feature Technology
5.1 Case Study: Antenna Parameter Tuning
5.1.1 Problem Statement
5.1.2 Response Features
5.1.3 Feature-Based Antenna Optimization
5.1.4 Illustration Example 1: Planar Inverted-F Antenna (PIFA)
5.1.5 Single-Brick Dielectric Resonator Antenna
5.1.6 Discussion
5.2 Case Study: Feature-Based Optimization of Dual-Band Coupler
5.2.1 Compact Dual-Band Microstrip Coupler: Geometry and Design Task
5.2.2 Response Features
5.2.3 Objective Function and Constraints: Optimization Algorithm
5.2.4 Numerical Results and Experimental Validation
5.3 Summary
References
Chapter 6: Response Features for Local Optimization
6.1 Conventional Feature-Based Optimization
6.1.1 Matching Enhancement of Multi-band Antennas
6.1.1.1 Design Optimization of High-Frequency Structures: Problem Statement
6.1.1.2 Feature Points Definition and Feature-Based Optimization Task
6.1.1.3 Trust Region Gradient-Based Search
6.1.1.4 Dual-Band Planar Antenna: Results
6.1.1.5 Triple-Band Antenna: Results
6.1.2 Feature-Based Optimization of an Impedance Matching Transformer
6.1.2.1 Bandwidth Maximization
6.1.2.2 Size Reduction
6.1.3 Feature-Based Optimization of a SIW Filter
6.1.3.1 FBO of SIW Filter: Results
6.1.4 FBO for the Design of Linear Microstrip Antenna Array Apertures
6.1.4.1 Setup of the Studies
6.1.4.2 Results
6.1.5 Conventional Feature-Based Optimization: Summary
6.2 FBO with Variable Number of Features
6.2.1 Response Features for UWB Antennas
6.2.2 Case Studies
6.2.2.1 Case I: Uniplanar UWB Antenna with CPW Feed
6.2.2.2 Case II: UWB Antenna with Rectangular Radiator
6.3 FBO with Circuit Decomposition
6.3.1 Conventional and Compact Circuit. Structure Decomposition
6.3.2 Cell Optimization Using Response Features
6.3.3 Simulation-Based Transformer Tuning
6.3.4 Optimization Algorithm
6.3.5 Results
6.4 Feature-Based Size Reduction
6.4.1 Problem Formulation
6.4.2 Controlling Reflection Response Using Response Features
6.4.3 Objective Function for FBO Design
6.4.4 Results
6.4.4.1 Dual-Band Quasi-Patch Antenna
6.4.4.2 Triple-Band Uniplanar Dipole Antenna
6.4.4.3 Experimental Validation
6.5 Summary
References
Chapter 7: Accelerated Feature-Based Local Optimization with Variable-Fidelity EM Simulations
7.1 Variable-Fidelity Feature-Based Optimization of Antennas
7.1.1 Problem Formulation
7.1.2 Automated Low-Fidelity Model Selection
7.1.3 Bandwidth Enhancement of a Quasi-Isotropic DRA
7.1.4 Matching Improvement of a Slot-Ring Coupled Patch Antenna
7.2 Variable-Fidelity Feature-Based Optimization of Microwave Structures
7.2.1 Response Features for Multi-Band Couplers
7.2.2 Design Objectives
7.2.3 Automated Low-Fidelity Model Selection
7.2.4 Results for Compact Dual-Band Coupler I
7.2.5 Results for Compact Dual-Band Coupler II
7.3 Summary
References
Chapter 8: Expedited Feature-Based Local Optimization: Adjoint Sensitivities and Sparse Sensitivity Updates
8.1 Feature-Based Optimization with Adjoint Sensitivities
8.1.1 Problem Formulation
8.1.2 Optimization Algorithm with Adjoint Sensitivities
8.1.3 Case Studies
8.1.3.1 Four-Aperture Iris Coupled Bandpass Filter
8.1.3.2 Compact Microstrip Rat-Race Coupler
8.2 Feature-Based Optimization with Sparse Sensitivity Updates
8.2.1 Globalized Feature-Based Optimization with Sparse Sensitivity Updates
8.2.1.1 Feature-Based Optimization of Multi-Band Antennas
8.2.1.2 Trust-Region Search with Jacobian Changes Tracking
8.2.2 Verification Case Studies
8.3 Summary
References
Chapter 9: Response Features for Reliability Improvement of Local Optimization Procedures
9.1 Design Specification Adaptation for Reliability Improvement
9.1.1 Design Specification Adaptation: Concept and Prerequisites
9.1.1.1 Specification Adjustment Concept
9.1.1.2 Specification Adjustment Prerequisites
9.1.1.3 Specification Adjustment Procedure
9.1.2 Optimization Engine: Trust-Region Algorithm
9.1.3 Optimization Algorithm
9.1.4 Results: Dual-Band Branch-Line Coupler
9.1.5 Results: Dual-Band Power Divider
9.2 Design Specification Management with Variable-Fidelity Models
9.2.1 Problem Formulation: Multi-Band Antennas
9.2.2 Improving Computational Efficiency: Variable-Resolution Models
9.2.2.1 Variable-Resolution Models
9.2.2.2 Model Resolution and Design Targets
9.2.2.3 Model Resolution and Convergence Status
9.2.3 Complete Optimization Framework
9.2.4 Results: Dual-Band Dipole
9.2.5 Results: Triple-Band Dipole
9.3 Feature-Based Frequency-Based Regularization
9.3.1 Optimization by Frequency-Based Regularization
9.3.2 Results: Quasi-Yagi Planar Antenna
9.3.3 Results: Ultra Wide-Band Antenna
9.3.4 Summary
9.4 Regularization with Multi-Fidelity Models and Sparse Sensitivity Updates
9.4.1 Restricted Sensitivity Updates
9.4.2 Optimization Framework
9.4.3 Results
9.4.3.1 Test Antenna Structures
9.4.3.2 Experimental Setup and Results
9.4.3.3 Discussion
9.5 Summary
References
Chapter 10: Response Features for Global and Multi-objective Optimization
10.1 Global Optimization Using Feature-Based Inverse Surrogates
10.1.1 Design Problem Formulation
10.1.2 Response Features for Initial Design Quality Assessment
10.1.3 Global Search Using Inverse Surrogates
10.1.4 Local Refinement by Trust-Region Gradient Search
10.1.5 Optimization Framework
10.1.6 Demonstration Case Study I: Triple-Band Antenna
10.1.7 Demonstration Case Study II: Rat-Race Coupler
10.2 Global Optimization Using Simplex-Based Regression Surrogates
10.2.1 Problem Formulation
10.2.2 Simplex-Based Predictors
10.2.3 Global Search by Means of Simplex-Based Predictors
10.2.3.1 Design Quality Assessment
10.2.3.2 Global Search: Automated Simplex Update
10.2.3.3 Objective Function Improvement
10.2.4 Local Tuning Using Gradient-Based Trust-Region Search
10.2.5 Optimization Procedure
10.2.6 Verification Case I: Quasi-Yagi Antenna
10.2.7 Verification Case II: Power Divider
10.3 Multi-objective Optimization with Features
10.3.1 Multi-objective Design Problem Formulation
10.3.2 Optimization Algorithm
10.3.3 Pareto Front Exploration Using Response Features
10.3.4 Verification Case: Ultra-wideband Monopole Antenna
10.4 Summary
References
Chapter 11: Response Features for Warm-Start Optimization and Rapid Database Acquisition
11.1 Expedited Parameter Tuning Using Inverse/Forward Surrogates and Response Features
11.1.1 Problem Formulation
11.1.2 Response Features
11.1.3 Inverse Model: Feature-Based Forward Surrogate-Initial Design
11.1.4 Design Refinement by Fast Gradient Search
11.1.5 Complete Procedure
11.1.6 Demonstration Examples-Case I: Ring-Slot Antenna
11.1.7 Demonstration Examples-Case II: Quasi-Yagi Antenna
11.1.8 Demonstration Examples-Case III: Rat-Race Coupler
11.2 Computationally-Efficient Reference Design Acquisition for Reduced-Cost Constrained Modeling and Redesign of Compact Micr...
11.2.1 Design Optimality: Database Design Acquisition Problem
11.2.2 Assessing Design Quality Using Response Features
11.2.3 Inverse Sensitivity
11.2.4 Rapid Database Design Acquisition by Inverse Sensitivity and Response Features
11.2.4.1 Initial Design Rendition Using Inverse Metamodels
11.2.4.2 Design Refinement Using Response Features
11.2.4.3 Optimization Algorithm
11.2.5 Demonstration Example: Miniaturized Branch-Line Coupler
11.3 Expedited Microwave Design Closure Using Kriging Surrogates and Iterative Correction
11.3.1 Preexisting Designs and Kriging Surrogates
11.3.2 Initial Design
11.3.3 Basic Design Refinement: Fast Gradient Search
11.3.4 Iterative Design Refinement
11.3.5 Demonstration Case Study: Three-Section Impedance-Matching Transformer
11.4 Automated Inverse Design of Bandpass Filters Through Linear Approximation of Physical Dimensions and Response Features
11.4.1 Filter Optimization Using Response Features
11.4.2 Coupling Coefficients as Filter Features
11.4.3 Filter Implementation Using Inverse Model of Design Parameters
11.4.4 Illustration and Application Example
References
Chapter 12: Response Features for Low-Cost Behavioral Modeling
12.1 Feature-Based Modeling Using Characteristic Point Allocation
12.1.1 Modeling Methodology
12.1.2 Case Study I: Dielectric Resonator Antenna
12.1.3 Case Study II: Patch Antenna Coupled to Rectangular Slot
12.1.4 Case Study III: Dual-Band Uniplanar Dipole Antenna
12.2 Feature-Based Modeling of Microwave Circuits
12.2.1 Modeling Methodology
12.2.2 Demonstration Case Studies
12.2.3 Application Case Studies
12.3 Variable-Fidelity Feature-Based Modeling
12.3.1 Modeling Procedure
12.3.1.1 Variable-Fidelity Feature-Based Modeling: The Internal Approach
12.3.1.2 Variable-Fidelity Feature-Based Modeling: The External Approach
12.3.1.3 Internal Versus External Approach
12.3.2 Verification Examples
References
Chapter 13: Constrained Modeling with Response Features
13.1 Feature-Based Nested Kriging
13.1.1 Nested Kriging
13.1.2 Nested Kriging with Response Features
13.1.3 Verification Case: Dual-Band Antenna
13.2 Feature-Based Nested Kriging with Dimensionality Reduction
13.2.1 Nested Kriging with Explicit Dimensionality Reduction
13.2.1.1 Domain Definition
13.2.2 PCA-Based Constrained Modeling with Response Features
13.2.2.1 Design of Experiments. Surrogate Model Optimization
13.2.3 Verification Case: Triple-Band Antenna
13.3 Two-Stage Observable-Based Constrained Modeling
13.3.1 Inverse Regression Surrogate
13.3.2 Domain Definition and Final Surrogate
13.3.3 Complete Two-Stage Observable-Based Modeling Framework
13.3.4 Verification Case Study I: Quasi-Yagi Antenna
13.3.5 Verification Case Study II: Compact Branch-Line Coupler
13.4 Two-Stage Feature-Based Modeling
13.4.1 Verification Case Study I: Ring-Slot Antenna
13.4.2 Verification Case Study II: Quasi-Yagi Antenna
13.5 Summary
References
Chapter 14: Feature-Based Uncertainty Quantification
14.1 Motivation
14.2 Fabrication Yield and Design Centering: Problem Formulation
14.3 Yield Optimization Using Response Features
14.3.1 Yield Optimization Task Using Response Features
14.3.2 Design Centering by Feature-Based Surrogates and Trust Regions
14.3.3 Demonstration Examples
14.4 Yield Optimization of Microwave Filters Using Feature-Based Linear Regression Models
14.4.1 Feature-Based Approximation Model
14.4.2 Yield Estimation Examples
14.4.3 Yield Optimization Examples
14.5 Expedited Feature-Based Uncertainty Quantification Using Variable-Fidelity EM Simulations
14.5.1 Variable-Fidelity Yield Optimization Using Second-Order Feature-Based Surrogates
14.5.2 Variable-Fidelity Yield Optimization Using Linear Feature-Based Surrogates and Trust Regions
14.6 Feature-Based Tolerance Optimization
14.6.1 Feature-Based Regression Model
14.6.2 Evaluating Objective Function: Tolerance Optimization Algorithm
14.6.3 Demonstration Examples 1: Antenna Design
14.6.4 Demonstration Examples 1: Microwave Design
14.7 Design Centering Using Response Features and Inverse Regression
14.7.1 Task Formulation and Models
14.7.2 Design Centering Algorithm
14.7.3 Demonstration Case Studies
14.8 Summary
References
Chapter 15: Generalized Response Features
15.1 Motivation
15.2 Response Features: Recollection of the Concept and Advantages
15.3 Generalized Response Features
15.3.1 Concept and Feature Point Definition
15.3.2 Objective Function Definition
15.3.2.1 Objective Function for In-Band Matching Improvement
15.3.2.2 Objective Function for Bandwidth Maximization
15.3.3 Optimization Algorithm Flow
15.4 Illustration Examples
15.4.1 Case I: Ring Slot Antenna
15.4.2 Case II: Triple-Band Dipole Antenna
15.4.3 Case III: Bandwidth-Enhanced Patch Antenna
15.4.4 Case IV: Ultra-Wideband Antenna
15.4.5 Discussion
15.5 Summary
References
Chapter 16: Summary
Index